1. Field of Invention
The present invention relates generally to the field of mobile communications. More particularly, in one exemplary aspect, the present invention is directed to synchronizing one network element's timing (e.g., slave) with that of another (e.g., master) network or network element.
2. Description of Related Technology
Many network elements in both wired and wireless networks needs to have accurate timing and frequency information. Time, phase, or frequency need to be accurate within the margin of error allowed by the system. Some systems require only frequency to be synchronized. Alternatively, some systems require only time to be synchronized. Some systems require both time and frequency to be synchronized. Yet further, some systems are not synchronized at all, but accurate timing and/or frequency information are needed on both sides of the communications devices for the system to function properly.
The aforementioned network elements can physically be located apart from one another. Also, they can be powered on and off independently. The oscillators on the network elements are not shared. In this environment, synchronizing the time/phase and frequency of the system within a designated margin or error requires some communication between the elements being synchronized.
This synchronization is required to facilitate many tasks and applications. For example, for two devices on the network to make simultaneous measurements, the time has to be synchronized. Such a situation exists commonly in test equipment.
Also, to hand off from one network to another network or one station to another station, frequency and time need to be synchronized. To receive and transmit data correctly across the network, the frequency reference used has to be accurate to a prescribed tolerance. Similarly, to reduce the buffering requirements on a streaming application, time needs to be synchronized. To prevent a loss of a frame in a streaming application, the frequency needs to be synchronized. Many other examples will be recognized by those of ordinary skill.
In the particular case of a cellular network, the user equipment (UE) and the basestation need to be frequency-aligned in order to communicate with each other. The UE typically locks its frequency to that of the basestation(s). Basestations get their frequency reference from the network, from a GPS receiver, or in some cases a highly accurate but expensive oscillator.
Furthermore, the user device or terminal (e.g., cellular phone) needs to be able to “hand-oft” from one basestation to another; i.e., migrate from one geographic coverage area associated with a first base station to another associated with a second base station. To aid this hand-off, the user terminal needs to measure the signal of the neighboring basestation to which it will be migrating. To facilitate this measurement, the timing of the basestation(s) can be aligned so as to shorten the time window that the user terminal needs to measure for each basestation. Also, when handing off to another basestation, the terminal needs to be frequency-aligned to the new (receiving) basestation so that it can lock into the RF signal of that basestation. In this example, both time and frequency synchronization are important.
Furthermore, it is noted that different cellular standards have different tolerance and requirements for the accuracy of frequency and time. CDMA (i.e. IS-95 and IS-2000) require both time and frequency to be accurate. GSM requires frequency to be accurate, but has a less stringent requirement for time.
Network based synchronization techniques (e.g. IEEE-1588, T1, E1, etc.) use the network (e.g. Ethernet or GPIB) to synchronize time and/or frequency. One such exemplary standard—IEEE-1588, provides sub-microsecond synchronization of real-time clocks in components of the network. While such network synchronization techniques are simple and relatively inexpensive to implement, they are typically intended for relatively localized system. Network based synchronization techniques can work on larger systems with larger errors and/or a longer synchronization time.
An atomic-based oscillator (e.g., Cesium or other atomic clock) is possibly the most accurate means to keep time. First, the network elements are synchronized, and subsequently the atomic-based oscillator is used to maintain the system time. For example, GPS satellites use this method to keep time synchronization. Depending on the acceleration experienced by the network element, compensation for acceleration must be made. The most salient disadvantage of this method is the great expense.
GPS (Global Positioning System) provides for a very accurate, sub-nanosecond time reference. Also, GPS provides geographic location data. Though the prices of GPS solutions in the past have been high, recent advancements in chipsets (integrated circuits) have brought the price of GPS suites down to a moderate, though not inexpensive, price point. However, the comparatively low link budget (loosely translated—radiated power) for the GPS system does not allow for the signal to reach with any certainty indoor locations, or those where obstructions are present. At a minimum, the network element using the GPS system has to be able to “see” four (4) GPS satellites in order to obtain a fix, though a more accurate measurement can be made if it can “see” more satellites within the constellation. Many cellular basestations use GPS to keep the time and frequency synchronized.
In another approach, television signals may be used to get an accurate timing and frequency reference. Much like the GPS system, a system using television signals can also provide location. One such implementation is described in U.S. Pat. No. 6,839,024. Unlike GPS, TV signals penetrate further into buildings and structures. In addition to the TV station, a location and timing servers are required in the system. Furthermore, monitor stations that are located via GPS are required. Also, a TV phase center database is required. Basically, the user terminal locates itself, and determines the time in a manner similar to GPS. It measures the pseudo-ranges of the stations, and uses a database to determine the location and uses that information to calculate time. Due to the system-level components required, there is significant additional cost and complexity associated with this solution.
As yet another approach, direct connectivity may be used; i.e., where the system directly connects the elements needing the synchronization. There are many ways to accomplish this synchronization, including e.g., via a SERDES (serializer-deserializer) device, which may decode embedded data to provide for clock recovery. This is often used on an equipment rack or other such installations. However, this method generally cannot be used in a public network since there cannot be a direct connection between the elements in such applications.
Network elements (i.e., any networked object) often need to have either timing or frequency lock/synchronization/accuracy. The network elements can lock to a central source (i.e. Master reference), or to each other. An element can be frequency locked, yet not time locked. In this case, there will be a DC offset in time. An element can also be time locked, yet not frequency locked. In this case, the time will drift. Frequency error compounds to time errors.
If the frequency error averages zero (0) over long periods of time, the time error will be bounded. However, if there is a bias in the frequency error, the time error will grow unbounded, so there must be a mechanism to adjust the time in addition to frequency.
Many applications need to have this type of synchronization. One such example is in a cellular network, where a user terminal moves between multiple basestations. Another example is in the case of measurements made by physically disparate devices such as test equipment. Yet another application comprises streaming applications such as those used for video or multimedia. All three of the foregoing examples have different system needs for time and/or frequency synchronization.
Accordingly, improved methods and apparatus are needed for providing accurate clocking (timing) and/or frequency information over a public network, that also are cost effective and efficient to manufacture and use. Ideally such improved apparatus and methods would require no significant modification or additions to extant infrastructure or communications protocols, and would allow for a great degree of flexibility in terms of operational implementation (e.g., could be used with various physical device configurations, network topologies, and communications protocols/air interfaces). It would also make maximal use of COTS (off-the-shelf) or commodity components so as to further reduce cost.